Downsampling genomic sequence data

1. A computer-implemented method for automatically downsampling DNA sequence data using variational autoencoders and preserving genomic integrity of an original FASTQ file, the method comprising:

processing the original FASTQ file by tagging one or more reads with a unique identifier and mapping the reads to a sample reference sequence;

executing, by an encoder, bootstrapping on genomic sequence data to produce resamples based on the original FASTQ file, wherein the encoder utilizes a probabilistic resampling framework to select a subset of reads from the original FASTQ file and performs optimization of bootstrapping to produce a set of resamples for a size;

assessing, by the encoder, unrepresentativeness and self-inconsistency of the resamples and selecting a representative resample according to the assessment;

building, by the encoder, a first vector representation from genotype likelihoods based on the selected representative resample;

generating by the encoder, biological information, wherein generating the biological information including mapping positional information and genotype likelihoods comprises:

aligning, by the encoder, a plurality of reads containing nucleotides to different locations on a sample genome;

generating, by the encoder, the mapping positional information for the nucleotides in the aligned plurality of reads, wherein the mapping positional information comprises: mapping positions, mapping scores and nucleotide quality scores of a plurality of reads that map; and

encoding, by the encoder, the mapping positions, the mapping scores and the nucleotide quality scores of a plurality of reads that map to a specific location of the sample genome into a second vector representation;

integrating, by an analytics engine, mapping positional information and the genotype likelihoods to identify an optimum vector representation of a resample; and

decoding, by a decoder, the identified optimum vector representation of the resample to obtain a down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file, wherein the decoder produces a final set of read fragments that are representative of the original FASTQ file, and wherein the decoder reconstructs the first vector representation that encodes the reads as they move through the pipeline to generate the down-sampled read file.

2. The computer-implemented method of claim 1 further comprising:

resampling, by the encoder, DNA fragments generated from massively parallel sequencing, without compromising the genomic integrity of the original FASTQ file.

3. The computer-implemented method of claim 1 further comprising:

executing, by a modified variational encoder, probabilistic resampling to obtain a second set of resamples containing a plurality of reads.

4. The computer-implemented method of claim 1 further comprising:

receiving, by the decoder, the representative resample as input; and producing, by the decoder, a reconstruction of a plurality of original input reads based on received representative resample.

5. The computer-implemented method of claim 1 further comprising:

outputting the down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file to a user.

6. The computer-implemented method of claim 1 further comprising:

generating, by the encoder, a file containing positional and genotype information.

7. The computer-implemented method of claim 1 , wherein generating the biological information further comprises:

obtaining, by the encoder, genotype likelihoods for one or more genomic locations from the plurality of reads, wherein the genotype likelihoods are a plurality of alleles with specified positional change on the plurality of reads.

8. A computer program product for automatically downsampling DNA sequence data using variational autoencoders and preserving genomic integrity of an original FASTQ file, the computer program product comprising:

one or more computer readable storage medium and program instructions stored on the one or more computer readable storage medium, the stored program instructions comprising:

program instructions to process the original FASTQ file by tagging one or more reads with a unique identifier and mapping the reads to a sample reference sequence;

program instructions to execute, by an encoder, bootstrapping on genomic sequence data to produce resamples based on the original FASTQ file, wherein the encoder utilizes a probabilistic resampling framework to select a subset of reads from the original FASTQ file and performs optimization of bootstrapping to produce a set of resamples for a size;

program instructions to assess, by the encoder, unrepresentativeness and self-inconsistency of the resamples and selecting a representative resample according to the assessment;

program instructions to build, by the encoder, a first vector representation from genotype likelihoods based on the selected representative resample;

program instructions to generate by the encoder, biological information, wherein generating the biological information including mapping positional information and genotype likelihoods comprises:

program instructions to align, by the encoder, a plurality of reads containing nucleotides to different locations on a sample genome;

program instructions to generate, by the encoder, the mapping positional information for the nucleotides in the aligned plurality of reads, wherein the mapping positional information comprises: mapping positions, mapping scores and nucleotide quality scores of a plurality of reads that map; and

program instructions to encode, by the encoder, the mapping positions, the mapping scores and the nucleotide quality scores of a plurality of reads that map to a specific location of the sample genome into a second vector representation;

program instructions to integrate, by an analytics engine, mapping positional information and the genotype likelihoods to identify an optimum vector representation of a resample; and

program instructions to decode, by a decoder, the identified optimum vector representation of the resample to obtain a down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file, wherein the decoder produces a final set of read fragments that are representative of the original FASTQ file, and wherein the decoder reconstructs the first vector representation that encodes the reads as they move through the pipeline to generate the down-sampled read file.

9. The computer program product of claim 8 further comprising:

program instructions to resample, by the encoder, DNA fragments generated from massively parallel sequencing, without compromising the genomic integrity of the original FASTQ file.

10. The computer program product of claim 8 further comprising:

program instructions to execute, by a modified variational encoder, probabilistic resampling to obtain a second set of resamples containing a plurality of reads.

11. The computer program product of claim 8 further comprising:

program instructions to receive, by the decoder, the representative resample as input; and

program instructions to produce, by the decoder, a reconstruction of a plurality of original input reads based on received representative resample.

12. The computer program product of claim 8 further comprising:

program instructions to output the down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file to a user.

13. The computer program product of claim 8 further comprising:

program instructions to generate, by the encoder, a file containing positional and genotype information.

14. The computer program product of claim 8 , wherein generating the biological information further comprises:

program instructions to obtain, by the encoder, genotype likelihoods for one or more genomic locations from the plurality of reads, wherein the genotype likelihoods are a plurality of alleles with specified positional change on the plurality of reads.

15. A computer system for automatically downsampling DNA sequence data using variational autoencoders and preserving genomic integrity of an original FASTQ file, the computer system comprising:

one or more computer processors;

one or more computer readable storage medium;

program instructions stored on the one or more computer readable storage medium for execution by at least one of the one or more computer processors, the stored program instructions comprising:

program instructions to process the original FASTQ file by tagging one or more reads with a unique identifier and mapping the reads to a sample reference sequence;

program instructions to execute, by an encoder, bootstrapping on genomic sequence data to produce resamples based on the original FASTQ file, wherein the encoder utilizes a probabilistic resampling framework to select a subset of reads from the original FASTQ file and performs optimization of bootstrapping to produce a set of resamples for a size;

program instructions to assess, by the encoder, unrepresentativeness and self-inconsistency of the resamples and selecting a representative resample according to the assessment;

program instructions to build, by the encoder, a first vector representation from genotype likelihoods based on the selected representative resample;

program instructions to generate by the encoder, biological information, wherein generating the biological information including mapping positional information and genotype likelihoods comprises:

program instructions to align, by the encoder, a plurality of reads containing nucleotides to different locations on a sample genome;

program instructions to generate, by the encoder, the mapping positional information for the nucleotides in the aligned plurality of reads, wherein the mapping positional information comprises: mapping positions, mapping scores and nucleotide quality scores of a plurality of reads that map; and

program instructions to encode, by the encoder, the mapping positions, the mapping scores and the nucleotide quality scores of a plurality of reads that map to a specific location of the sample genome into a second vector representation;

program instructions to integrate, by an analytics engine, mapping positional information and the genotype likelihoods to identify an optimum vector representation of a resample; and

program instructions to decode, by a decoder, the identified optimum vector representation of the resample to obtain a down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file, wherein the decoder produces a final set of read fragments that are representative of the original FASTQ file, and wherein the decoder reconstructs the first vector representation that encodes the reads as they move through the pipeline to generate the down-sampled read file.

16. The computer system of claim 15 further comprising:

program instructions to resample, by the encoder, DNA fragments generated from massively parallel sequencing, without compromising the genomic integrity of the original FASTQ file.

17. The computer system of claim 15 further comprising:

program instructions to execute, by a modified variational encoder, probabilistic resampling to obtain a second set of resamples containing a plurality of reads.

18. The computer system of claim 15 further comprising:

program instructions to receive, by the decoder, the representative resample as input; and

program instructions to produce, by the decoder, a reconstruction of a plurality of original input reads based on received representative resample.

19. The computer system of claim 15 further comprising:

program instructions to output the down-sampled read file that resembles and maintains the genomic integrity of the original FASTQ file to a user; and

program instructions to generate, by the encoder, a file containing positional and genotype information.

20. The computer system of claim 15 , wherein generating the biological information further comprises:

program instructions to obtain, by the encoder, genotype likelihoods for one or more genomic locations from the plurality of reads, wherein the genotype likelihoods are a plurality of alleles with specified positional change on the plurality of reads.